MRI Mapping With Multi-Submodel AI for Shorter Single-Scan Acquisition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing T1, T2, and T1rho mapping techniques in magnetic resonance imaging suffer from long scan times and motion artifacts, requiring separate scans and breath-holding, which increases patient discomfort and decreases imaging efficiency.
Innovation Solution
A method utilizing a trained machine learning model with multiple sub-models to process MR images, enabling simultaneous generation of T1, T2, and T1rho mapping images in a single scan, reducing the number of required images and scan time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate scans are performed for T1, T2, and T1rho mapping, then mapping accuracy is improved, but scan time increases
Solution Approach 1:
The patent combines multiple separate mapping scans (T1, T2, T1rho) into a single integrated scan sequence. The MRI scanner performs all three mapping acquisitions simultaneously by interleaving the pulse sequences, allowing the system to obtain all mapping data in one go rather than requiring three separate scans, thus reducing total scan time while maintaining mapping accuracy
Solution Approach 2:
The patent develops a unified processing framework that can handle multiple types of mapping data (T1, T2, T1rho) through a single scan sequence. The machine learning model is trained to process diverse mapping types using a common architecture, enabling the system to perform multiple mapping functions simultaneously without requiring separate dedicated sequences for each mapping type
2Measurement precision
If extended scan duration is used for accurate mapping, then mapping quality is improved, but motion artifacts increase
Solution Approach 1:
By merging multiple mapping acquisitions into a single scan sequence, the patent ensures that all mapping data is collected within one breath-hold period. This eliminates the problem of patient motion between separate scans, as the entire mapping process completes before the patient needs to resume breathing, thereby reducing motion artifacts while maintaining mapping quality
Solution Approach 2:
The patent performs all necessary mapping data acquisition within a single breath-hold period before any motion can occur. By completing the entire mapping sequence in advance, the system captures all required data while the patient is still in the same physiological state, preventing motion-related degradation of mapping quality
3Loss of information
If multiple separate scans are required, then comprehensive mapping data is obtained, but patient discomfort increases
Solution Approach 1:
The patent merges T1, T2, and T1rho mapping acquisitions into a single scan that can be completed within one breath-hold. This eliminates the need for patients to perform multiple separate breath-holds, reducing discomfort while ensuring all mapping data is comprehensively captured in one continuous acquisition
Solution Approach 2:
The patent collects all necessary mapping data during a single breath-hold period before the patient needs to resume breathing. By completing all data acquisition in advance, the system ensures comprehensive mapping information is obtained while minimizing the total time the patient must hold their breath, thereby improving comfort
4Loss of information
If multiple separate scans are performed, then complete mapping information is acquired, but imaging efficiency decreases
Solution Approach 1:
The patent combines multiple mapping sequences into a single integrated scan that acquires T1, T2, and T1rho data simultaneously. This merging approach ensures complete mapping information is obtained while reducing the total scan time from what would be required for three separate scans, thereby improving imaging efficiency without sacrificing information completeness
Solution Approach 2:
The patent implements a universal processing framework that handles multiple mapping types through a single scan sequence and unified machine learning model. This multi-functional approach allows the system to efficiently process diverse mapping data in one go, improving throughput and imaging efficiency while maintaining complete mapping information acquisition
Data Source
AI summary
Embodiments of the present disclosure provides a method implemented on a computing device including at least one processor and a storage device. The method, may include obtaining magnetic resonance (MR) images of a subject, at least two of the MR images being acquired by an MRI scanner according to different imaging parameters. The method may also include obtaining one or more target MR mappings corresponding to at least a portion of the MR images by processing the MR images through a trained machine learning model. The second count of the target MR mappings may be less than a first count of the MR images. The trained machine learning model may include at least two sub-models, and each sub-model processes at least one of the MR images.